Large-scale Conformation Research Articles

A large-scale conformation sampling and evaluation server for protein tertiary structure prediction and its assessment in CASP11.

BackgroundWith more and more protein sequences produced in the genomic era, predicting protein structures from sequences becomes very important for elucidating the molecular details and functions of these proteins for biomedical research. Traditional template-based protein structure prediction methods tend to focus on identifying the best templates, generating the best alignments, and applying the best energy function to rank models, which often cannot achieve the best performance because of the difficulty of obtaining best templates, alignments, and models.MethodsWe developed a large-scale conformation sampling and evaluation method and its servers to improve the reliability and robustness of protein structure prediction. In the first step, our method used a variety of alignment methods to sample relevant and complementary templates and to generate alternative and diverse target-template alignments, used a template and alignment combination protocol to combine alignments, and used template-based and template-free modeling methods to generate a pool of conformations for a target protein. In the second step, it used a large number of protein model quality assessment methods to evaluate and rank the models in the protein model pool, in conjunction with an exception handling strategy to deal with any additional failure in model ranking.ResultsThe method was implemented as two protein structure prediction servers: MULTICOM-CONSTRUCT and MULTICOM-CLUSTER that participated in the 11th Critical Assessment of Techniques for Protein Structure Prediction (CASP11) in 2014. The two servers were ranked among the best 10 server predictors.ConclusionsThe good performance of our servers in CASP11 demonstrates the effectiveness and robustness of the large-scale conformation sampling and evaluation. The MULTICOM server is available at: http://sysbio.rnet.missouri.edu/multicom_cluster/.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-015-0775-x) contains supplementary material, which is available to authorized users.

Physical origin of the large-scale conformity in the specific star formation rates of galaxies

Two explanations have been put forward to explain the observed conformity between the colours and specific star formation rates (SFR/$M_*$) of galaxies on large scales: 1) the formation times of their surrounding dark matter halos are correlated (commonly referred to as "assembly bias"), 2) gas is heated over large scales at early times, leading to coherent modulation of cooling and star formation between well-separated galaxies (commonly referred to as "pre-heating") . To distinguish between the pre-heating and assembly bias scenarios, we search for relics of energetic feedback events in the neighbourhood of central galaxies with different specific star formation rates. We find a significant excess of very high mass ($\log M_* > 11.3$) galaxies out to a distance of 2.5 Mpc around low SFR/$M_*$ central galaxies compared to control samples of higher SFR/$M_*$ central galaxies with the same stellar mass and redshift. We also find that very massive galaxies in the neighbourhood of low SFR/$M_*$ galaxies have much higher probability of hosting radio loud active galactic nuclei. The radio-loud AGN fraction in neighbours with $\log M_* > 11.3$ is four times higher around passive, non star-forming centrals at projected distances of 1 Mpc and two times higher at projected distances of 4 Mpc. Finally, we carry out an investigation of conformity effects in the recently publicly-released Illustris cosmological hydrodynamical simulation, which includes energetic input both from quasars and from radio mode accretion onto black holes. We do not find conformity effects of comparable amplitude on large scales in the simulations and we propose that gas needs to be pushed out of dark matter halos more efficiently at high redshifts.

Shear-stress-induced conformational changes of von Willebrand factor in a water-glycerol mixture observed with single molecule microscopy.

The von Willebrand factor (VWF) is a human plasma protein that plays a key role in the initiation of the formation of thrombi under high shear stress in both normal and pathological situations. It is believed that VWF undergoes a conformational transition from a compacted, globular to an extended form at high shear stress. In this paper, we develop and employ an approach to visualize the large-scale conformation of VWF in a (pressure-driven) Poiseuille flow of water-glycerol buffers with wide-field single molecule fluorescence microscopy as a function of shear stress. Comparison of the imaging results for VWF with the results of a control with λ-phage double-stranded DNA shows that the detection of individual VWF multimers in flow is feasible. A small fraction of VWF multimers are observed as visibly extended along one axis up to lengths of 2.0 μm at high applied shear stresses. The size of this fraction of molecules seems to exhibit an apparent dependency on shear stress. We further demonstrate that the obtained results are independent of the charge of the fluorophore used to label VWF. The obtained results support the hypothesis of the conformational extension of VWF in shear flow.

Biomolecular Recognition Based on 3D Molecular Theory of Solvation

Protein-protein and protein-ligand recognition plays an essential role in many biomolecular processes. Understanding the mechanisms of protein-ligand binding is also crucial for designing new and optimizing existing therapeutic agents, and in different biotechnological applications. It is well known that solvation effects, including hydrophobicity and solvent mediated hydrogen bonding, are major factors implicated in biomolecular processes. Such processes often involve slow conformational changes and exchange of solvent and ligand molecules between bulk solution and biomolecular cavities.We develop new approaches for accurate account of molecular solvation effects in protein-ligand recognition and binding, which further extend the ligand mapping protocols based on the molecular theory of solvation, a.k.a. the three-dimensional reference interaction site model with the Kovalenko-Hirata closure (3D-RISM-KH). Based on statistical mechanics, this theory provides a natural link between different levels of coarse-graining details in a multiscale description of solvation structure and thermodynamics, from highly localized structural solvent and bound ligand molecules to effective desolvation potentials and self-assembling nanoarchitectures in solvents of different composition in a range of thermodynamic conditions.Implemented in the new 3D-RISM-Dock protocol, the theory provides a quantitative estimate of binding affinities, based on a detailed physical description of hydrogen bonding, hydrophobicity, and solvation entropic effects, with full account for molecular specificities, thermodynamic conditions, and concentration effects. The theory accurately predicts the solvation structure of biomolecules. This allows us to incorporate the 3D-RISM-KH description of structural solvation in a new docking protocol. The latter has been successfully applied to predict the binding modes of the periplasmic binding proteins in situations when structural solvation and desolvation effects are of importance. We also show that the 3D-RISM-KH theory provides valuable information on the role of the solvation entropic effects in the ligand binding and large scale conformation dynamics of proteins.

Allosteric Regulation of Protein Kinase Enzymes via an Electrostatic Switch that Modulates Active Site Dynamics

Protein kinase (PK) enzymes are a large family of signaling proteins that play a central role in signal transduction pathways. Robust regulation of their catalytic activity is critical, and many oncogenes harbor mutations that result in misregulated PK activity. The chemical basis for how some PK regulatory factors ultimately affect the rate of chemistry is still not completely understood. We have identified a long-range electrostatic switch that we believe is used by allosteric PK regulatory factors to modulate the rate of chemistry by tuning active-site dynamics.We applied a combination of crystallography, kinetics, and molecular dynamics to determine the chemical kinetic basis for how this electrostatic switch, toggled by regulatory subunit binding, affects each step of catalysis by CDK2 kinase. We engineered point mutants to deconstruct the kinetic, dynamic, and thermodynamic consequences of the switch. We also evaluated other PKs and find that, although it has evolved to be triggered in different ways by diverse PK regulatory factors, the mechanics of this switch can be conserved.We demonstrate that a key component of the switch is that it affects a significant change in the electrostatic potential within the ATP∗Mg binding site of the enzyme. This electrostatic effect is propagated through the low-dielectric protein interior and directly affects the two dominant rate-determining steps of catalysis: attenuating both the recruitment of catalytically essential Mg cofactors (affecting both kcat and KM) as well as the release of the ADP product.Conclusion: We present a chemical hypothesis that provides a mechanistic explanation for one way that a large-scale conformation transition, observed in diverse PK family members, is able to significantly affect the rate of chemistry by acting at a distance from the active site.

Seafloor control on sea ice

The seafloor has a profound role in Arctic Sea ice formation and seasonal evolution. Ocean bathymetry controls the distribution and mixing of warm and cold waters, which may originate from different sources, thereby dictating the pattern of sea ice on the ocean surface. Sea ice dynamics, forced by surface winds, are also guided by seafloor features in preferential directions. Here, satellite mapping of sea ice together with buoy measurements are used to reveal the bathymetric control on sea ice growth and dynamics. Bathymetric effects on sea ice formation are clearly observed in the conformity between sea ice patterns and bathymetric characteristics in the peripheral seas. Beyond local features, bathymetric control appears over extensive regions of the sea ice cover across the Arctic Ocean. The large-scale conformity between bathymetry and patterns of different synoptic sea ice classes, including seasonal and perennial sea ice, is identified. An implication of the bathymetric influence is that the maximum extent of the total sea ice cover is relatively stable, as observed by scatterometer data in the decade of the 2000s, while the minimum ice extent has decreased drastically. Because of the geologic control, the sea ice cover can expand only as far as it reaches the seashore, the continental shelf break, or other pronounced bathymetric features in the peripheral seas. Since the seafloor does not change significantly for decades or centuries, sea ice patterns can be recurrent around certain bathymetric features, which, once identified, may help improve short-term forecast, seasonal outlook, and decadal prediction of the sea ice cover. Moreover, the seafloor can indirectly influence the cloud cover by its control on sea ice distribution, which differentially modulates the latent heat flux through ice covered and open water areas.

Correlated motion of protein subdomains and large-scale conformational flexibility of RecA protein filament

Based on X-ray crystallographic data available at Protein Data Bank, we have built molecular dynamics (MD) models of homologous recombinases RecA from E. coli and D. radiodurans. Functional form of RecA enzyme, which is known to be a long helical filament, was approximated by a trimer, simulated in periodic water box. The MD trajectories were analyzed in terms of large-scale conformational motions that could be detectable by neutron and X-ray scattering techniques. The analysis revealed that large-scale RecA monomer dynamics can be described in terms of relative motions of 7 subdomains. Motion of C-terminal domain was the major contributor to the overall dynamics of protein. Principal component analysis (PCA) of the MD trajectories in the atom coordinate space showed that rotation of C-domain is correlated with the conformational changes in the central domain and N-terminal domain, that forms the monomer-monomer interface. Thus, even though C-terminal domain is relatively far from the interface, its orientation is correlated with large-scale filament conformation. PCA of the trajectories in the main chain dihedral angle coordinate space implicates a co-existence of a several different large-scale conformations of the modeled trimer. In order to clarify the relationship of independent domain orientation with large-scale filament conformation, we have performed analysis of independent domain motion and its implications on the filament geometry.

Use of Fast DNA Footprinting and Real-Time Fluorescence Spectroscopy to Characterize Novel, Biologically Revlevant Transcription Initiation Intermediates for E. Coli RNA Polymerase

RNA polymerase (RNAP) is a complex molecular motor responsible for catalyzing the synthesis of RNA from DNA in a promoter sequence-dependent manner. Knowledge of the mechanism of transcription initiation is crucial to understanding the regulation of gene expression. However, the RNAP-DNA intermediates on the pathway to open complex formation during transcription initiation are extremely short-lived and thus have been very difficult to characterize using traditional structural methods. Here, we present work focused on developing rapid, solution-based methods that allow us to characterize RNAP-DNA structures in the millisecond time regime. These include: fast DNA footprinting (using hydroxyl radical and permanganate probes) to characterize the interaction of RNAP with the DNA backbone and the extent of thymine base unstacking; stopped-flow FRET experiments to monitor DNA bending in real time; and 2-aminopurine fluorescence experiments to characterize the kinetics of DNA opening. In a related project, we utilize thin layer chromatography and fluorescence spectroscopy to characterize the short (2, 3 and 4-mer) RNA products made by RNAP during repeated rounds of abortive cycling before the enzyme progresses to forming full-length RNA transcripts. These combinatorial approaches have allowed us to determine the timing of loading of duplex DNA into the RNAP active site cleft and the kinetic mechanism of large-scale RNAP conformation changes leading to initial RNA synthesis. These research activities have successfully developed new methods to characterize transient intermediates in solution and to begin to develop a detailed structural-mechanistic picture of transcription initiation.

Replica Exchange with Solute Scaling: A More Efficient Version of Replica Exchange with Solute Tempering (REST2)

A small change in the Hamiltonian scaling in Replica Exchange with Solute Tempering (REST) is found to improve its sampling efficiency greatly, especially for the sampling of aqueous protein solutions in which there are large-scale solute conformation changes. Like the original REST (REST1), the new version (which we call REST2) also bypasses the poor scaling with system size of the standard Temperature Replica Exchange Method (TREM), reducing the number of replicas (parallel processes) from what must be used in TREM. This reduction is accomplished by deforming the Hamiltonian function for each replica in such a way that the acceptance probability for the exchange of replica configurations does not depend on the number of explicit water molecules in the system. For proof of concept, REST2 is compared with TREM and with REST1 for the folding of the trpcage and β-hairpin in water. The comparisons confirm that REST2 greatly reduces the number of CPUs required by regular replica exchange and greatly increases the sampling efficiency over REST1. This method reduces the CPU time required for calculating thermodynamic averages and for the ab initio folding of proteins in explicit water.

BAC TG-EMBED: one-step method for high-level, copy-number-dependent, position-independent transgene expression

Chromosome position effects combined with transgene silencing of multi-copy plasmid insertions lead to highly variable and usually quite low expression levels of mini-genes integrated into mammalian chromosomes. Together, these effects greatly complicate obtaining high-level expression of therapeutic proteins in mammalian cells or reproducible expression of individual or multiple transgenes. Here, we report a simple, one-step procedure for obtaining high-level, reproducible mini-gene expression in mammalian cells. By inserting mini-genes at different locations within a BAC containing the DHFR housekeeping gene locus, we obtain copy-number-dependent, position-independent expression with chromosomal insertions of one to several hundred BAC copies. These multi-copy DHFR BAC insertions adopt similar large-scale chromatin conformations independent of their chromosome integration site, including insertions within centromeric heterochromatin. Prevention of chromosome position effects, therefore, may be the result of embedding the mini-gene within the BAC-specific large-scale chromatin structure. The expression of reporter mini-genes can be stably maintained during continuous, long-term culture in the presence of drug selection. Finally, we show that this method is extendable to reproducible, high-level expression of multiple mini-genes, providing improved expression of both single and multiple transgenes.

Insights into Interphase Large-Scale Chromatin Structure from Analysis of Engineered Chromosome Regions

How chromatin folds into mitotic and interphase chromosomes has remained a difficult question for many years. We have used three generations of engineered chromosome regions as a means of visualizing specific chromosome regions in live cells and cells fixed under conditions that preserve large-scale chromatin structure. Our results confirm the existence of large-scale chromatin domains and fibers formed by the folding of 10-nm and 30-nm chromatin fibers into larger, spatially distinct domains. Transcription at levels within severalfold of the levels measured for endogenous loci occur within these large-scale chromatin structures on a condensed template linearly compacted several hundred fold to 1000-fold relative to B-form DNA. However, transcriptional induction is accompanied by a severalfold decondensation of this large-scale chromatin structure that propagates hundreds of kilobases beyond the induced gene. Examination of engineered chromosome regions in mouse embryonic stem cells (ESCs) and differentiated cells suggests a surprising degree of plasticity in this large-scale chromatin structure, allowing long-range DNA interactions within the context of large-scale chromatin fibers. Recapitulation of gene-specific differences in large-scale chromatin conformation and nuclear positioning using these engineered chromosome regions will facilitate identification of cis and trans determinants of interphase chromosome architecture.

The Small Angle Scattering Toolbox

Small Angle Scattering (SAS) is a technique used to investigate structure and dynamics of macromolecules in solution. Proteins in buffer conditions close to their physiological environment, are subject to Xray or Neutron scattering experiments. The resulting one-dimensional scattering curves are directly related to their three-dimensional structure. The SAS technique is routinely used to determining the low resolution shape of protein and map specific large scale conformation changes in protein structures.We present a recently developed computational platform for SAS data analysis and model construction/refinement.The Small Angle Scattering Toolbox (SASTBX) has tools four major modules: (1) Raw data reduction; (2) theoretical scattering profile calculation based on PDB structures; (3) Pair distance distribution function (PDDF) estimation; and (4) 3D model construction and structure refinement.The toolbox can be utilized to read raw scattering images obtained from the detector to generate an intensity profile. The basic analyses, such as Guiner and Kratky plots can be carried out in real time to assess the sample and data quality while collecting data. The PDDF estimation is a fully automated procedure, linked with a database a known PDDF's allowing for a rough initial classification of the shape of the protein. Model data can be calculated on the basis of a spherical harmonics expansions. Initial structures can be further refined with normal mode movements or rigid-body motions.The sastbx is built on the open source Computational Crystallography Toolbox (CCTBX). The toolbox is implemented by using Python/C++ hybrid approach: the computing intensive jobs are handled in C++, and the python allows easy integration between other components. The source code will be distributed as open source project.

Modeling G protein‐coupled receptors for structure‐based drug discovery using low‐frequency normal modes for refinement of homology models: Application to H3 antagonists

G Protein-Coupled Receptors (GPCRs) are integral membrane proteins that play important role in regulating key physiological functions, and are targets of about 50% of all recently launched drugs. High-resolution experimental structures are available only for very few GPCRs. As a result, structure-based drug design efforts for GPCRs continue to rely on in silico modeling, which is considered to be an extremely difficult task especially for these receptors. Here, we describe Gmodel, a novel approach for building 3D atomic models of GPCRs using a normal mode-based refinement of homology models. Gmodel uses a small set of relevant low-frequency vibrational modes derived from Random Elastic Network model to efficiently sample the large-scale receptor conformation changes and generate an ensemble of alternative models. These are used to assemble receptor-ligand complexes by docking a known active into each of the alternative models. Each of these is next filtered using restraints derived from known mutation and binding affinity data and is refined in the presence of the active ligand. In this study, Gmodel was applied to generate models of the antagonist form of histamine 3 (H3) receptor. The validity of this novel modeling approach is demonstrated by performing virtual screening (using the refined models) that consistently produces highly enriched hit lists. The models are further validated by analyzing the available SAR related to classical H3 antagonists, and are found to be in good agreement with the available experimental data, thus providing novel insights into the receptor-ligand interactions.

Peeling the Onion: Ribosomes Are Ancient Molecular Fossils

We describe a method to establish chronologies of ancient ribosomal evolution. The method uses structure-based and sequence-based comparison of the large subunits (LSUs) of Haloarcula marismortui and Thermus thermophilus. These are the highest resolution ribosome structures available and represent disparate regions of the evolutionary tree. We have sectioned the superimposed LSUs into concentric shells, like an onion, using the site of peptidyl transfer as the origin (the PT-origin). This spherical approximation combined with a shell-by-shell comparison captures significant information along the evolutionary time line revealing, for example, that sequence and conformational similarity of the 23S rRNAs are greatest near the PT-origin and diverge smoothly with distance from it. The results suggest that the conformation and interactions of both RNA and protein can be described as changing, in an observable manner, over evolutionary time. The tendency of macromolecules to assume regular secondary structural elements such as A-form helices with Watson-Crick base pairs (RNA) and alpha-helices and beta-sheets (protein) is low at early time points but increases as time progresses. The conformations of ribosomal protein components near the PT-origin suggest that they may be molecular fossils of the peptide ancestors of ribosomal proteins. Their abbreviated length may have proscribed formation of secondary structure, which is indeed nearly absent from the region of the LSU nearest the PT-origin. Formation and evolution of the early PT center may have involved Mg(2+)-mediated assembly of at least partially single-stranded RNA oligomers or polymers. As one moves from center to periphery, proteins appear to replace magnesium ions. The LSU is known to have undergone large-scale conformation changes upon assembly. The T. thermophilus LSU analyzed here is part of a fully assembled ribosome, whereas the H. marismortui LSU analyzed here is dissociated from other ribosomal components. Large-scale conformational differences in the 23S rRNAs are evident from superimposition and prevent structural alignment of some portions of the rRNAs, including the L1 stalk.

Protein conformational transitions explored by mixed elastic network models

We develop a mixed elastic network model (MENM) to study large-scale conformational transitions of proteins between two (or more) known structures. Elastic network potentials for the beginning and end states of a transition are combined, in effect, by adding their respective partition functions. The resulting effective MENM energy function smoothly interpolates between the original surfaces, and retains the beginning and end structures as local minima. Saddle points, transition paths, potentials of mean force, and partition functions can be found efficiently by largely analytic methods. To characterize the protein motions during a conformational transition, we follow "transition paths" on the MENM surface that connect the beginning and end structures and are invariant to parameterizations of the model and the mathematical form of the mixing scheme. As illustrations of the general formalism, we study large-scale conformation changes of the motor proteins KIF1A kinesin and myosin II. We generate possible transition paths for these two proteins that reveal details of their conformational motions. The MENM formalism is computationally efficient and generally applicable even for large protein systems that undergo highly collective structural changes.

DNA molecules on periodically microstructured lipid membranes: Localization and coil stretching

We explore large scale conformations of DNA molecules adsorbed on curved surfaces. For that purpose, we investigate the behavior of DNA adsorbed on periodically shaped cationic lipid membranes. These unique membrane morphologies are supported on grooved, one-dimensionally periodic microstructured surfaces. Strikingly, we find that these periodically structured membranes are capable to stretch DNA coils. We elucidate this phenomenon in terms of surface curvature dependent potential energy attained by the adsorbed DNA molecules. Due to it, DNA molecules undergo a localization transition causing them to stretch by binding to highly curved sections (edges) of the supported membranes. This effect provides a new venue for controlling conformations of semiflexible polymers such as DNA by employing their interactions with specially designed biocompatible surfaces. We report the first experimental observation of semiflexible polymers unbinding transition in which DNA molecules unbind from one-dimensional manifolds (edges) while remaining bound to two-dimensional manifolds (cationic membranes).

Elasticity ofα-Helical Coiled Coils

Predicting large scale conformations of protein structures is computationally demanding. Here we compute the conformation and elasticity of double-stranded coiled coils using a simple coarse-grained elastic model. By maximizing the contact between hydrophobic residues and minimizing the elastic energy, we show that the minimum energy structure of a coiled coil is a supercoiled double helix of alpha helices. For realistic binding energies, the elastic energy of the alpha helices requires binding every 7th residue, which leads to a pitch and helix angle for the structure that is consistent with experimental measurements. Analysis of the model equations shows how the pitch varies with the helical repeat of the hydrophobic residues and with the ratio of the twisting modulus to the bending modulus and provides an estimate of the persistence length of around 150 nm, in agreement with previous experimental estimates.

Thermal, Structural, and Viscoelastic Characterization of cis-Poly(phenylene vinylene) Related to its Photo-Isomerization

A cis-poly(phenylene vinylene) (cis-PPV) recently synthesized by Katayama, Ozawa, and coworkers undergoes photo-isomerization into trans configuration. This isomerization occurred rather quickly within 45 min in solvent-cast films at room temperature. The cis-PPV chains exhibiting this quick isomerization were expected to be highly mobile in the bulk state, and this expectation was examined from thermal, structural, and viscoelastic tests for a cis-PPV sample with Mn=4300 and Mw/Mn=2.16. It turned out that the as-polymerized cis-PPV was partially crystalline and fully melted at Tm≅109 °C. The samples quenched from a molten state to lower temperatures (≤70°C<Tm) as well as cast from a solution were amorphous for few hours, enabling the viscoelastic test in the fully amorphous state. The WLF-type time-temperature superposition was valid for these quenched and/or cast sample, confirming the amorphous character of the cis-PPV chains therein. The terminal viscoelastic relaxation time of these chains, equivalent to a time required for thermal changes in a large-scale conformation, was ≅10 s at room temperature. This result suggested that the conformational changes of the amorphous chains (the major component in the cast films) were sufficiently fast to allow the efficient cis-to-trans photo-isomerization.

An enhanced elastic network model to represent the motions of domain‐swapped proteins

Domain swapping is a process where two (or more) protein molecules form a dimer (or higher oligomer) by exchanging an identical domain. In this article, based on the observation that domains are rigid and hinge loops are highly flexible, we propose a new Elastic Network Model, domain-ENM, for domain-swapped proteins. In this model, the rigidity of domains is taken into account by using a larger spring constant for intradomain contacts. The large-scale transition of domain swapping is then novelly decomposed into the relative motion between the rigid domains (only 6 degrees of freedom) plus the internal fluctuations of each domain. Consequently, this approach has the potential to produce much more meaningful transition pathways than other simulation approaches that try to find pathways in a search space of large numbers of dimensions. In this article, we also propose a new way to define the overlap measure. Past approaches used an inappropriate comparison of the large-scale conformation displacement against the computed infinitesimal motions of modes. Here, we propose an infinitesimal version of the large-scale conformation change and then compare it with the modes of motions. As a result, we obtain much better overlap values. Using this new overlap definition, we are also able for the first time to give a clear, intuitive explanation why "open" forms tend to produce better overlap values than "closed" forms with traditional ENMs. Finally, as an application, we present a simple approach to show how domain-ENM can be used to generated transition pathways for domain-swapped proteins.

A protein folding pathway with multiple folding intermediates at atomic resolution.

Using native-state hydrogen-exchange-directed protein engineering and multidimensional NMR, we determined the high-resolution structure (rms deviation, 1.1 angstroms) for an intermediate of the four-helix bundle protein: Rd-apocytochrome b562. The intermediate has the N-terminal helix and a part of the C-terminal helix unfolded. In earlier studies, we also solved the structures of two other folding intermediates for the same protein: one with the N-terminal helix alone unfolded and the other with a reorganized hydrophobic core. Together, these structures provide a description of a protein folding pathway with multiple intermediates at atomic resolution. The two general features for the intermediates are (i) native-like backbone topology and (ii) nonnative side-chain interactions. These results have implications for important issues in protein folding studies, including large-scale conformation search, -value analysis, and computer simulations.